Multi-Touch Attribution and Marketing Mix Modelling
Statistical and survey-based methods for estimating the real contribution of search and AI visibility when click-level tracking no longer joins the journey together.
Analogy Mental model
Many customers saw a recommendation in a chat assistant or walked past a poster in a lift, then came in and bought, and no camera recorded any of it. So you ask at the till how they heard about you, and you compare months when the poster spend went up against the months that followed.
In practice What you actually do
As AI answers absorb discovery, a growing share of buyers never click the citation and instead search your brand name later. Add one question at checkout or signup asking how they found you, with explicit options for search, an AI assistant, video, and community forums. Self-reported attribution closes a gap that no tracking pixel can.
Where people get burned Pitfalls
The question has to be placed where completion is high and the options have to name the channels you actually want to measure. A vague free-text field produces anecdotes, while a short list produces a dataset you can put in a board deck.
This entry was written in English from the Chinese original, not machine-translated. Read the Chinese version →